Ecosystem effects of the Atlantic Multidecadal Oscillation

Journal of Marine Systems 133 (2014) 103–116

Contents lists available at ScienceDirect

Journal of Marine Systems

j ourna l homepage: www.e lsev ie r .com/ locate / jmarsys

Ecosystem effects of the Atlantic Multidecadal Oscillation

Janet A. Nye a,⁎, Matthew R. Baker b, Richard Bell c, Andrew Kenny d, K. Halimeda Kilbourne e,Kevin D. Friedland c, Edward Martino f, Megan M. Stachura g, Kyle S. Van Houtan h,i, Robert Wood f

a United States Environmental Protection Agency, Office of Research and Development, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division,27 Tarzwell Dr, Narragansett, RI 02882, USAb NOAA National Marine Fisheries Service, Alaska Fisheries Science Center, 7600 Sand Point Way N.E., Seattle, WA 98115, USAc NOAA National Marine Fisheries Service, Northeast Fisheries Science Center, 28 Tarzwell Dr., Narragansett, RI 02882, USAd Centre for Environment, Fisheries & Aquaculture Science, Lowestoft, UKe Chesapeake Biological Laboratory, Center for Environmental Science, University of Maryland, Box 38, Solomons, MD 20688, USAf National Oceanic and Atmospheric Administration, Cooperative Oxford Laboratory, Oxford, MD 21654, USAg School of Aquatic and Fishery Sciences, University of Washington, Box 355020, Seattle, WA 98195–5020, USAh NOAA Fisheries Service, Pacific Islands Fisheries Science Center, Honolulu, HI 96822, USAi Nicholas School of the Environment and Earth Sciences, Duke University, Durham, NC 27708, USA

⁎ Corresponding author at: School of Marine and AtmUniversity, Stony Brook, NY 11794-5000, USA. Tel.: +1

E-mail address: [email protected] (J.A. Nye

0924-7963/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.jmarsys.2013.02.006

a b s t r a c t

Article history:Received 26 April 2012Received in revised form 21 November 2012Accepted 15 February 2013Available online 21 February 2013

Keywords:Climate variabilityClimate changeEcosystem based management

Multidecadal variability in the Atlantic Ocean and its importance to the Earth's climate system has been the sub-ject of study in the physical oceanography field for decades. Only recently, however, has the importance of thisvariability, termed the Atlantic Multidecadal Oscillation or AMO, been recognized by ecologists as an importantfactor influencing ecosystem state. A growing body of literature suggests that AMO-related fluctuations are asso-ciated with shifts in ecological boundaries, primary productivity, and a number of ecologically and economicallyimportant coastal and marine populations across the Atlantic basin. Although the AMO is a basin-wide index ofSST, the drivers of ecosystem change encompass more than temperature anomalies and the mode of action dif-ferswithin each ecosystem. A common theme in assessing ecosystem change indicates that fluctuations inwatermasses and circulation patterns drive shifts in ecosystem states, but the magnitude and rate of change is depen-dent on the physical characteristics of the region. Because of thewide ranging geographic effects of the AMO, andconsidering its multidecadal nature, a more complete understanding of its causes and effects would allow scien-tists and managers to more effectively inform ecosystem-based management across the Atlantic Basin.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

There are several modes of climatic variability in the North Atlanticthat affect ecosystem processes, but one mode of variability that has re-ceived relatively little attention until recently is the AtlanticMultidecadalOscillation (AMO sensu Kerr, 2005) also known as Atlantic MultidecadalVariability (AMV sensu Delworth et al., 2007; Knight et al., 2005). It ishypothesized that fluctuations in the strength of Atlantic MeridionalOverturning Circulation (AMOC) cause internal variability in sea surfacetemperature (SST), sea level pressure, and ocean circulation all of whichare represented by the AMO index (Knight et al., 2005; Ting et al., 2011).Many biological oceanographers are familiar with the AMO index andrefer to this large-scale phenomenon as the AMO. For consistency wewill use this terminology throughout this review of its effect on ecosys-tems. However, the AMO index represents a wide variety of processessuch that AMV is perhaps the more appropriate terminology for this

ospheric Sciences, Stony Brook401 782 3165.).

rights reserved.

phenomenon. In particular, records of past climate variability indicatethat the AMO is not an oscillatory cycle with regular periods of fixedlength (Gray et al., 2004; Knudsen et al., 2011). Instead it appears to bea climate system feature with variance concentrated at multidecadalscales.

The AMO index (Fig. 1) is typically defined as the SST anomalyfrom 0-60oN linearly detrended to account for the increase in temper-ature associated with anthropogenic climate change (Enfield et al.,2001; Sutton and Hodson, 2005). Modern observations of SST indicatethat the AMO switches between positive and negative phases on theorder of 65–70 years (Schlesinger and Ramankutty, 1994), but the lengthand consistency of the oscillatory cycle is the subject of considerable de-bate. The 65–70 year cycle is based on only ~130 years of observed andreconstructed SST data for which there are only 1.5-2 complete cyclesof the AMO. Smoothing or detrending of SST to calculate the AMOindex results in oscillations of different frequencies (Vincze and Janosi,2011). Although the exact timing of the switch from the positive to neg-ative phase depends on how the index is calculated, it is generally agreedthat negative/cold phases occurred from approximately 1900–1925 and1971–1994 while positive/warm phases occurred from 1875–1899,1926–1970 and 1990–present (Goldenberg et al., 2001).

Fig. 1. The unsmoothed and smoothed AMO index calculated from Kaplan SST dataset detrended for the effects of climate change (Enfield et al., 2001). The dashed line is theunsmoothed, undetrended SST data with the 1951–1980 NOAA ERSST climatology added back in. All data obtained from the NOAA ESRL website (http://www.esrl.noaa.gov/psd/data/timeseries/AMO/).

104 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

The large scale warming and cooling related to the AMO interactswith anthropogenic warming. It has been suggested that the combinedeffects of anthropogenic climate change and the positive phase of theAMO since the 1990s has caused a more rapid warming than wouldbe expected from climate change alone (Andronova and Schlesinger,2000; Belkin, 2009; Knudsen et al., 2011). Similarly, cool (negative)phases of the AMO in the past may have masked the effects of climatechange. The combination of warming trends from AMO and fromanthropogenic climate change since the 1970s makes it difficult to dis-tinguish the cause of changes in ecological time series unless the recordlength extends back before the mid to late 20th century.

The purpose of this review is to first explain the physical phenomenaassociated with the AMO so as to elucidate how this broad scale climaticindex may be associated with more localized ecosystems in the Atlanticbasin. Because the AMO index is usually presented as a time series in theecological literature, we focus on the spatial aspects of the AMO to betterunderstand the more proximate mechanisms by which this large scaleprocess affects local ecosystemdynamics. Secondly,we reviewpublishedecological studies where the AMO was found to influence populationsand ecosystems orwherewe suspect that the AMOmay have influencedecosystem dynamics. Lastly, we will discuss how the AMO affects largemarine ecosystems (LMEs) around the Atlantic and how understandingthis process is fundamental for informing Ecosystem Based approachesto Management (EBM).

2. Spatial patterns of the AMO

In contrast to the time series of the AMO index, the spatially explicitrepresentation of the AMO suggests that the mechanisms throughwhich this phenomenon affects ecosystems varies in different areas ofthe Atlantic and perhaps in other parts of the globe (Fig. 2). As in previ-ous studies (Delworth et al., 2007; Grossmann and Klotzbach, 2009), ahorseshoe-shaped spatial pattern (warm colors in Fig. 2) is generatedin the Atlantic when the AMO is correlated with SST in the positivephase of the AMO. Thus, the regional effects of the AMO can varythroughout the Atlantic basin, but the positive phase of the AMO gener-ally indicates a period of warmer temperatures. Whereas the AMO isprimarily defined by oceanic phenomena (SST anomalies), it is relatedto atmospheric processes aswell, since the ocean and atmosphere inter-act closely to form the Earth's climate system. During the positive phaseof the AMO the position of the Intertropical Convergence Zone (ITCZ)

shifts from the south (Fig. 2, brown ellipse), where precipitation isreduced, to the north, where precipitation is increased (Fig. 2, greenellipse). This ITCZ shift is associated with weaker northeast tradewinds and a stronger cross equatorial wind flow from the southernequatorial zone. The weaker northern hemisphere winds are the resultof a weakening of the Bermuda High and Icelandic Low atmosphericpressure zones. The AMO also appears to generate remote effects withanomalously low pressure over Eastern Europe and perhaps also overthe northeast Pacific. The position of high and low pressure cells resultsin easterly winds over the central North Atlantic that influence the po-sition and strength of the Gulf Stream and North Atlantic current. Themixed layer depth (MLD) is also shallower in the positive phase. Thenegative phase of the AMO has roughly opposite sign anomalies, butgiven the very long time scale of the AMO, there is not enough data todetermine how linear the signal is (i.e. the extent to which the positiveand negative phases are equal and opposite).

While it may be tempting to assume from this schematic that theareas with highest spatial correlations are the areas where ecosystemchanges related to the AMO are most frequently found, the AMO'sinfluence is still very strong in areas where correlations are low. For ex-ample, the highest correlations occur to the east of Spain (Delworthet al., 2007), while low correlations occur near the Chesapeake Bay,Northwest Atlantic and the continental US. However, strong ecosystemeffects are observed in North America in part because of the influenceof SST on atmospheric processes such as precipitation and windpatterns.

3. Links between AMO and other modes of climate variability

In addition to the AMO, other patterns of climate variability interactwith the AMO to elicit ecosystem response. Such climatic processesinclude the North Atlantic Oscillation (NAO), Arctic Oscillation (AO) orNorthern Annular Mode (NAM), Atlantic Meridional Mode (AMM),El Niño and the Southern Oscillation (ENSO) and the Pacific DecadalOscillation (PDO). The NAO is a north–south dipole in sea level pressure(SLP) that is primarily governed by internal atmospheric dynamics,although it can be influenced by both local and nonlocal SST anomalies(Hurrell and Deser, 2009; Hurrell et al., 2003). It affects the ocean via anumber of processes, including surface heat fluxes, which drives a SSTtripole pattern in the North Atlantic and influences deep water forma-tion in the Labrador Sea, and the wind stress curl, which can alter the

Fig. 2. Schematic of climate perturbations associated with AMO warm phase. The shaded base map represents the correlation of annual SST anomalies from NCEP/NCAR reanalysis1948–2007 with the AMO index. Drier and wetter conditions are indicated by brown and green ellipses, respectively. Wind anomalies are represented by thick black arrows and ashallower ocean mixed layer in the North Atlantic Gyre is indicated by the text “MLD shoals”. Open blue and red ellipses represent low and high pressure regions, respectively. Basemap image provided by the NOAA/ESRL Physical Sciences Division, Boulder Colorado from their Web site at http://www.esrl.noaa.gov/psd/.

105J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

gyre circulations (Häkkinen and Rhines, 2004; Häkkinen et al., 2011).The NAO may influence the AMO through changes in sea ice dynamicsand North Atlantic Deep Water formation (Grossmann and Klotzbach,2009). Both of these processes influence thermohaline circulation andthus, the AMO. The AMO may in turn influence the NAO via changesin SST that affect pressure gradients (Grossmann and Klotzbach,2009). The AO is closely associated with the NAO, but is defined as thedominant pattern of non-seasonal sea-level pressure variations northof 20 N.

Two processes in the Pacific may relate to the Atlantic climate vari-ability through complex teleconnections, the extent to which is unclear(but seeMcKinnell paper, this volume). ENSO is centered in the tropicalPacific, but can impact the global ocean including the North Atlantic.The PDO, defined by the first Empirical Orthogonal Function (EOF) ofNorth Pacific SST, exhibits variability on multiple time scales with en-hanced variability at a period of ~20 years. The PDO is likely controlledby several factors including random fluctuations in the Aleutian Low,ENSO teleconnections, and coupled atmosphere–ocean interaction inthe North Pacific. Both the AMO and PDO exhibit multidecadal variabil-ity and while it appears that one may lag the other, no mechanism tolink the two currently exists (d'Orgeville and Peltier, 2007). However,both are influenced by Western Boundary Currents, the Kuroshio inthe Pacific (Kwon and Deser, 2007) and the Gulf Stream in the Atlantic(Frankignoul et al., 2001; Joyce et al., 2009). Thus, analogous internalocean dynamics may influence these basin-scale processes in similarways to give rise to multidecadal climate variability, but how the AMOand PDO directly influence each other remains unclear (Kwon et al.,2010).

The AMO is newer to the scientific community, but already has apresence in the scientific literature on par with other often cited indicesof climate variability. Fig. 3a shows the results of our survey of the ISIWeb of Science publications for the NAO, PDO, and AMO since theirfirst record in this database. Since 1987, the NAO has been the centraltopic of 3661 publications, and these studies have been cited 90,726times. The PDO, since 1997, has 785 publications that have been cited13,065 times. And since 2000, the AMO has 210 publications that havebeen cited on 2859 occasions. As the AMO is increasingly understoodand appreciated, its scientific importance as expressed in the literaturemay equal that of the NAO and the PDO. Because the AMO is an emerg-ing research topic, most AMO publications to date describe the indexitself and its relationship to local and regional climate. Far fewer studiesaddress its significance to biological populations and ecosystems.

4. Taxa-specific responses to AMO

There are several mechanisms by which the AMO may influencemarine species. Temperature affects all physiological processes oforganisms, especially ectotherms (Fry, 1971; Hoar, 1953). Thus, growth,consumption, metabolism, migration and reproductive output may beaffected by AMO-related temperature changes. Changes in temperaturealone may affect the population growth rates of lower trophic levels(phytoplankton and zooplankton) much more than upper trophiclevels, by reducing generation time under optimal temperature condi-tions, changing dormancy cycles for certain planktonic species, and al-tering phenology. These changes in lower trophic levels could cascadethrough the food web and fundamentally alter ecosystem state. Mobile

Fig. 3. Publication rate and topics researched for published AMO studies. Using the ISIWeb of Science database, we compare the publications and citations for theNorth AtlanticOscillation (NAO), Pacific Decadal Oscillation (PDO), and the AMO. (a.) After 25 yearsin the literature, roughly 300 NAO publications and 13,000 citations appear annually.Though not as mature in the scientific literature, PDO and AMO studies are steadily in-creasing on pace with the NAO. (b.) Most AMO studies (69%) address the index itselfand its relationship to regional climate indicators, such as rainfall or temperature. Lessthan one third (30%) of AMO publications directly relate the AMO to ecological populationmetrics, andmost of these publications (14%) investigate precipitation through the proxyseries of terrestrial vegetation records. The relationship of the AMO to terrestrial and ma-rine populations is still a young research field.

106 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

marine organisms may respond to temperature by moving away fromsuboptimal temperatures.

The AMO is also associated with changes in wind and currentregimes (Delworth et al., 2007; Häkkinen et al., 2011). Thus, passivedispersal of marine organisms, many of which are planktonic or haveplanktonic larvae, will be affected resulting in changes in spatial distri-bution, survival, and population abundance. The response of each spe-cies to the physical changes caused by the AMO will depend on lifehistory, spawning behavior, and temperature preferences/tolerances.These individual species responses to the AMO and their subsequentinteractions with other components in the ecosystem may lead tounexpected consequences.

The majority of AMO publications (144 or 69%) did not researchthe influence of the AMO on any taxonomic group, reflecting that byfar most AMO studies address the index itself and its relationship toclimate indicators, such as rainfall or temperature (Fig. 3b). Of the210 total AMO publications recognized through our ISI search, 36(17%) dealt with autotrophs, 15 (7%) with vertebrates, and 13 (6%)with invertebrates. Less than one third (64 or 30%) of AMO publica-tions directly relate the AMO to ecological population metrics, andmost of these publications (30 or 14%) investigate precipitationthrough the proxy series of terrestrial vegetation records. The rela-tionship of the AMO to terrestrial and marine populations is still adeveloping research field.

4.1. Phytoplankton

Shifts in global and basin-specific primary production and phyto-plankton abundances are related to large-scale climatic variability inthe physical environment. Despite the lack of a definitive periodassociated with the AMO, the relationship between the atmosphericvariability associated with the NAO and thermohaline circulationcharacterized by the AMO suggests a coupling of ocean and atmo-spheric processes on time scales that will facilitate understandingand prediction of inter-annual weather, precipitation, fluvial outputand estuarine and marine productivity. Several studies have docu-mented shifts related to the AMO in oceanic (Martinez et al., 2009),coastal and continental shelf (Rodrigues et al., 2009; Schofield et al.,2008) and estuarine (Hubeny et al., 2006) productivity and responsesin forage fish populations to this increase in planktonic productivity(Tourre et al., 2007).

Correlations between the AMO indices and marine productivityhave been documented at immediate (Schofield et al., 2008), historical(Martinez et al., 2009; Tourre et al., 2007), and pre-historical (Hubenyet al., 2006; Rodrigues et al., 2009) scales. There are several mechanismsthrough which this might occur. Stratification within the water columnseparates nutrients at lower depths from entering the euphotic zone,thus limiting primary productivity. Long-term climate oscillations mayaugment or suppress seasonal conditions that promote or erode stratifi-cation and regulate mixing. Schofield et al. (2008) have demonstratedlinks between the AMO and phytoplankton blooms in the Mid-AtlanticBight, such that increased nutrient availability during the annual falltransition were augmented by increased winter winds associated witha positive phase in the AMO. Martinez et al. (2009) demonstrated thatglobal phytoplankton productivity is linked to multi-decadal oscillationsin both the Atlantic and Pacific, suggesting that links between chloro-phyll and sea surface temperature are driven by interactions betweenthe pycnocline and mixed surface layers. The influence of large-scaleclimate fluctuations on surface stratification has particular influence onphytoplankton productivity in the tropics and mid-latitudes, waterswhich are often nutrient-limited. Alternatively, warmer sea surfacetemperatures themselves may influence productivity. Shifts in fluvialinputs associated with altered precipitation patterns will also alterriver-induced fertilization of marine productivity.

There are several means to evaluate correlations between marineproductivity and multi-decadal shifts in climate. Chlorophyll concen-trations and SST may be monitored through satellite observations andimaging. Photosynthetic pigments and organic compounds such asalkenones present in sediments provide historical time-series on pri-mary production among autotrophs. Alternatively proxies for primaryproductivity such as survey or landing indices of forage fish abun-dance or biomass may be indicators of changes in productivity.Multi-decadal low frequency signals in the populations of small pe-lagics and the relative synchrony in these populations across oceanbasins suggest populations respond not only to basin-wide climaticpatterns but to changes in lower trophic levels (Chavez et al., 2003;Tourre et al., 2007).

107J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

4.2. Zooplankton

Zooplankton is especially susceptible to oceanographic changesdriven by the AMO because of their short life-cycles and high dispersalrates within their bioclimatic envelope (Beaugrand et al., 2009). North-ward geographical shifts of northeast Atlantic calanoid copepods havebeen observed in response to changes in SST, positively correlatedwith theAMO (Beaugrand et al., 2009). However, analysis for several re-gions within the northeast Atlantic of the abundance and the timing ofpeak abundance of copepod groups showed varied relationships withSST and AMO by region, indicating region-specific responses shouldbe considered (McGinty et al., 2011). In the Mid-Atlantic of the USNortheast Continental Shelf, increasing abundance of certain zooplank-ton taxa was also positively correlated with the AMO and warm watertemperatures (Kane, 2011).

Influences of the AMO on zooplankton may have important ecosys-tem implications through indirect effects on higher trophic level organ-isms. Multidecadal variability of North American salmon abundancecorrelated with the AMO is hypothesized to be a result of influences onthe abundance, concentration, and location of key zooplankton species(Condron et al., 2005). Survival of Atlantic salmon in southern Europe iscorrelated with the AMO and the abundance of several prey species ofzooplankton (Friedland et al., 2009). The copepod Calanus finmarchicusis a key food item of cod and other organisms in the northeast Atlantic(Beaugrand et al., 2003). It is expected to decline in some regions withincreasing summer temperatures, due in part to the AMO, and the declinecould lead to broad ecosystem impacts (Kamenos, 2010).

4.3. Invertebrates

The AMO influences larger invertebrate species through both directand indirect effects. The observed decline of several demersal fish spe-cies in Narragansett Bay, Rhode Island from 1959 to 2005 in correlationwith the AMO has had consequences for other species in this ecosystem(Collie et al., 2008). Several of these demersal fish species feed primarilyon crustaceans and their decline may have triggered an increase in crab(Cancer irroratus, C. borealis) and lobster (Homarus americanus) speciesthrough a release from predation. Increases in squid (Loligo pealeii)positively correlated with the AMO also correspond to a shift from de-mersal to pelagic species, resulting from decreasing chlorophyll a andincreasing zooplankton grazing (Collie et al., 2008).Hydrological regimesimpacting juvenile blue crab (Callinectes sapidus) abundance in the northcentral Gulf of Mexico were related to the AMO (Sanchez-Rubio et al.,2011). High abundance of juvenile blue crab during the early part ofthe time series corresponded with a cold AMO and positive NAO phase.The abundance was low however during the later part of the time serieswhen the AMOhad changed to awarmphase and theNAOhad switchedto a negative phase (Sanchez-Rubio et al., 2011).

Assemblages of invertebrate rocky intertidal assemblages aroundthe British Isles have also changed over the last 100 years in responseto temperature. With 100 years of data, these changes may be associ-ated with climate change, but several patterns suggest the AMO con-tributes to these observations. Since the 1990s the abundance ofwarm water species has increased. For instance, the limpet Patelladepressa was at high abundance during a warm phase of the AMO inthe 1950s then declined during a cold AMO phase, and increased inabundance during the recent warm AMO phase (Hawkins et al.,2009).

Multidecadal variability may contribute to extreme conditions thatcan impact reef-building corals. For example, anomalously high seasurface temperatures in the Caribbean, partly attributed to high back-ground SSTs during the positive phase of the AMO, were responsiblefor a recent mass coral bleaching event (Simonti and Eastman, 2010).Additionally, the strong relationship between Atlantic tropical cyclonefrequency and AMO (Goldenberg et al., 2001), indicates that coralreefs are more at risk from storm damage during warm phases of

AMO, when tropical cyclones are more prevalent (Mendoza et al.,2006).

Historical reconstructions of the AMO behavior can be gatheredfrom several invertebrates such as the bivalve Arctica icelandica, forami-nifera, and several species of scleractinian coral. Oxygen isotope analy-sis of Arctica islandica was used to investigate historical Gulf of Mainewater temperatures (Wanamaker et al., 2008). Findings indicated theAMO accounts for about 6–19% of local water temperature variabilityfrom 1950 to 2003 (Wanamaker et al., 2008). The AMO impacts oceantemperature, salinity and potentially many other environmental pa-rameters, which can manifest as changes in growth rates or in skeletalchemistry (Kilbourne et al., in review).

4.4. Fish

On both sides of the Atlantic there have been large scale shifts in thedistribution of fish species over the last century (Nye et al., 2009; Perryet al., 2005; Simpson et al., 2011; Sundby and Nakken, 2008; terHofstede et al., 2010). Large-scale coherence among different speciesor different populations suggests external factors are important driversof population dynamics (Moran, 1953). Multidecadal climate oscilla-tions exhibit forcing on basinwide scales and can have importantimpacts on the distribution and abundance of marine fish (Collie et al.,2004, 2008). During positive phases of the AMO there is generallya poleward shift in the distributions of marine organisms and asubsequent equatorial shift during negative phases. Range shifts canbe curtailed depending on the dispersal potential of the species, theconnectivity of suitable habitat patches and the rate of temperaturechange (Genner et al., 2004; ter Hofstede et al., 2010). Several potentialmechanisms explaining observed shifts in fish spatial distributioninclude directedmovement to remainwithin preferred habitat, popula-tion level processes such as spatial differences in recruitment or ecosys-tem changes in productivity or trophic interactions.

The spawning intensity of Arcto-Norwegian cod off the coast ofNorway has covaried with the AMO over the last hundred years. Thenorthern spawning areas have been the most important during thewarm phases of the AMO (~1930–1950, ~1980–present) while thesouthern spawning areas were most important during the cold phases(~1900–1920,~1960s–1970). The stock is at the northern extent of itsrange and is thus thermally limited. High temperatures increase fecun-dity and juvenile growth rate resulting in higher recruitments andabundance. The higher fecundity in Arcto-Norwegian cod is linked totheir hepatosomatic index and their condition factor which are both afunction of feeding conditions. During positive phases of the AMO,warm, productive water is advected from the Norwegian Sea reducingice coverage. This copepod-rich water increases growth rates of codcreating conditions conducive for Arcto-Norwegian cod to recruit welland increase in abundance (Sundby and Nakken, 2008). Althoughthe AMO was not implicated specifically, long term fluctuations(1908–1998) in Norwegian spring spawning herring is positively corre-lated with winter water temperatures (Toresen and Østvedt, 2000).

Along the east coast of the United States most stocks captured in afishery independent trawl survey exhibited a shift in response to thewarming conditions associated with the AMO (Nye et al., 2009). In gen-eral, stocks at the southern edge of their range decreased in abundanceand their center of abundance shifted poleward. In contrast, stocks atthe northern edge of their range typically increased in abundance andtheir range expanded northward. Those species constrained by thelack of connected suitable habitat, such as those confined to the Gulfof Maine, shifted their biomass to deeper water over the course of thetime series. The distribution of fish stocks appeared to shift in order toremain within their preferred temperature range (Nye et al., 2009).

In the northeast Atlantic, distributional studies showed thatsouthern fishes moved north into the English Channel, Celtic Seaand North Sea and within the North Sea species moved polewardover the last few decades (Perry et al., 2005; Simpson et al., 2011;

108 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

ter Hofstede et al., 2010). Since 1993, both winter bottom tempera-tures and species richness have increased in the North and CelticSea. However, species richness off the west coast of Scotland has de-clined in response to the warmer temperatures (ter Hofstede et al.,2010). Lusitanian (warm-water) species have expanded their rangeinto the North and Celtic Seas and their abundance is correlatedwith a five year running mean of bottom temperature. Boreal specieshave showed no trend over the same time period resulting in an in-crease in species richness. The boreal species off the west coast ofScotland have declined however and Lusitanian species have notmoved into the area. The temperature has increased more rapidlythan expected in these areas leading to the assertion that the AMOhas intensified warming attributed to global warming. Fishing mor-tality has declined in the northeast Atlantic since 1993 suggestingthat the climate and not overfishing is playing a more dominantrole in patterning fish abundances (ter Hofstede et al., 2010). Europe-an anchovy and sardine, warmer water species, have shifted northand reinvaded the North Sea and adjacent seas (Alheit et al., 2012).The NAO allows sardines to close its life cycle in the North Seawhile the warmer summer temperatures associated with the AMOlead to anchovy spawning in the North Sea.

The AMO can also have a direct effect on the total biomass of fishspecies. Atlantic salmon stocks are doing poorly in the eastern andwestern Atlantic despite limited or no fishing over the past twodecades (Friedland et al., 2009; Friedland et al., in review). The lowfrequency signal of Atlantic salmon catch, a proxy for abundance,was significantly correlated with the AMO from 1917–2009. Catchwas higher during the two cool periods 1905–1925 and 1970–1990and lower during the bracketing warm periods. The mortality ofjuveniles during their first year at sea appears to be an important bot-tleneck for this species. In the northeast Atlantic, the AMO impactsthe marine food web supporting juvenile salmon, resulting in lowergrowth during the warm phase and lower survivorship that is as-sumed to be driven by size mediated mortality. The warm phase ofthe AMO is believed to affect northwest Atlantic salmon by a differentmechanism. Juvenile marine growth does not appear to play a role,instead AMO related warming modifies the predator field affectingthe mortality rate of salmon at ocean entry and during the early ma-rine phase. Climate also appears to affect the accumulation of adultbiomass by thermal related mechanisms (Friedland and Todd, 2012).

The AMO has been linked to recruitment in other fishes using timeseries methodologies. Gröger and Fogarty (2011) developed ARIMAXtime series models to predict recruitment in cod and found thatmodels incorporating the AMO index, the NAO index and SpawningStock Biomass (SSB) vastly outperformed all other models. Acrossthe Atlantic basin, the recruitment of North Sea herring was analyzedwith a similar method and the ARIMAX model again outperformedthe traditional stock recruitment models. The best fit model includedthe AMO and NAO, but did not include SSB. Variations in herringrecruitment appear to be better explained by climate than by SSB.While the exact mechanisms regulating recruitment are not alwaysknown, climate oscillations capture large scale oceanographic pro-cesses which can have major impacts on the early life stages of fish(Gröger et al., 2010).

Year class strength is typically determined during the first year oflife (Houde, 1987) and influenced by environment conditions, whichare known to be more variable at the edges of a species range(Grinnell, 1922; Myers, 1998). This recruitment variability is oftenrelated to a species ability to extend its range in a poleward oranti-poleward fashion (Hare and Able, 2007; Hare et al., 2010). Theproposed mechanisms affecting recruitment in marine fishes includedirect temperature effects and/or changes in food availability. Incoastal systems, rainfall has substantial impacts on river flow, coastalnutrients and salinity, which can regulate the scale and timing ofprimary production. Fishes are known to time spawning to coincidewith the spring bloom productivity to enhance cohort survival, thus

changes to bloom phenology may adversely affect recruitment(Cushing, 1969; Platt et al., 2003). Warmer temperatures can alsoalter the activity or spatial and temporal distribution of potentialpredators (Keller and Klein-MacPhee, 2000; Taylor and Collie,2003). Though the mechanisms are still undescribed for many spe-cies, the accumulated body of evidence points to the AMO as a regu-lator of fish biomass through bottom-up effects on recruitmentdynamics (Drinkwater, 2006, 2011; Gröger and Fogarty, 2011;Gröger et al., 2010).

4.5. Sea turtles, birds, and marine mammals

The effect of the AMO on populations of higher order vertebratesis poorly understood. Our lack of knowledge stems from a poor un-derstanding of the effect of the AMO on primary production and fishpopulations, which provide the forage for sea turtles, birds, and ma-rine mammals. Despite an absence of comprehensive information,progress on relating the AMO to higher vertebrates is being made. Apopulation model formulated with the AMO as an independent vari-able explained up to 88% of the observed variability in loggerheadsea turtle (Caretta caretta) nesting over the past several decades(Van Houtan and Halley, 2011). It was hypothesized that the AMOhad a controlling effect on juvenile recruitment in the North Atlantic.Like most taxa, hatchling and neonatal sea turtles have physiologicalconstraints (Bostrom et al., 2010; Prange, 1976) that limit their capac-ity to exploit and endure resource variability. As the AMO index ispositively correlated with atmospheric and thermohaline circulationin the subtropical-temperate North Atlantic (Knight et al., 2005),the study found a positive AMO was linked to enhanced juvenile re-cruitment (Van Houtan and Halley, 2011). These ideas were recentlyapplied to explain the historical population variability of loggerheadturtles in the Pacific Ocean and to forecast environmental baselines(based on decadal oscillations) and model proposed fishery actions(Van Houtan, 2011). Knowledge of how the AMO relates to oceano-graphic processes and marine ecology will advance the understand-ing of the long-term population dynamics in higher order marinetaxa.

5. Ecosystem level responses to the AMO

5.1. Terrestrial systems

The AMO is derived from the variable dynamics of the marineenvironment, yet some of the best examples of linkages betweenAMO and ecosystem dynamics come from terrestrial systems. Whenthe AMO was first described by Schlesinger and Ramankutty (1994)as a phenomena that occurs primarily in the North Atlantic Ocean,the description included correlations with effects in terrestrial eco-systems in North America, Eurasia, and Africa. Warm phases of theAMO have been associated with low precipitation, droughts, reducedstream flow and elevated temperature in the southern continental US(Enfield et al., 2001; Sutton and Hodson, 2005, 2007). Because tem-perature and precipitation greatly affect plant growth, many proxytime series for the AMO are derived from tree ring chronology(Gray et al., 2004). Indeed, ~47% of the citations about effects of theAMO on autotrophs are for terrestrial species (Fig. 2).

The warm phase of the AMO is also associated with droughts inMexico, wet conditions in Europe, Africa, and India, and frequencyand intensity of Atlantic hurricanes (Feng et al., 2011; Goldenberget al., 2001; Hu and Feng, 2008; Sutton and Hodson, 2005; Zhangand Delworth, 2006). Many studies of the AMO's effects on humanpopulations (Fig. 2), relate North American drought conditions tothe persistence of civilizations (Benson et al., 2007; Mendoza et al.,2006). Melting of glaciers in the Swiss Alps (Huss et al., 2010),snow cover in the Tibetan Plateau (Shen et al., 2011), and sea icecover in Hudson Bay (Tivy et al., 2010) are all correlated with the

109J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

AMO; though the changes in the physical environment were welldocumented in these studies, the consequences for associated terres-trial and aquatic ecosystems have seldom been documented.

Because temperature and precipitation greatly affect plant growth,the AMO has dramatic effects on both agriculture and plant growth inthe wild. Forest growth is correlated with the AMO to the extent thatmany of the proxy time series for the AMO are derived from tree ringchronology (Gray et al., 2004). Tourre et al. (2011) showed that theAMO influences growing season and harvest times of grapes by theviticulture industry in France. In a meta-analysis of AMO related publi-cations, we found that ~47% of the citations about effects of the AMO onautotrophs were for terrestrial species (Fig. 3b).

5.2. Estuarine and aquatic ecosystems

The AMO has a direct effect on estuarine ecosystemsmainly due toits linkage to precipitation patterns and resultant river flows and es-tuarine mixing dynamics. The positive phase of the AMO is generallyassociated with reduced precipitation across the continental UnitedStates although there are exceptions in the southeast and parts ofthe mid-Atlantic region where precipitation is positively correlatedwith the AMO (Enfield et al., 2001). Moisture reconstructions from afreeze core at a lake in New Jersey partly supports the positive associ-ation of AMO and wet climate patterns in the Mid‐Atlantic regionover the last ∼240 years (Zhao et al., 2010). Cronin et al. (2003)note salinity oscillations in the Chesapeake Bay on time scales of60–70 years. Further, lake level records derived from long cores re-vealed an association between US northeast moisture conditionsand Atlantic sea surface temperatures at the multidecadal scale forthe last ∼7000 years (Li et al., 2007). Water transparency and thuslight availability in Florida lakes was strongly correlated with AMO,the more proximate mechanism being the influence of the AMO onrainfall (Gaiser et al., 2009).

Estuarine fossil pigments indicate that productivity and runoff werelower in a Northeast US estuary when the AMOwas in a cold, dry phase(Hubeny et al., 2006). Striped bass abundance increased while shadabundance declined in the Hudson River Estuary during positiveAMO time periods (Buchsbaum and Powell, 2008; O'Connor, 2010).Chesapeake Bay anadramous fish recruitments have been linked tointer-annual variability in hydroclimate conditions and associatedeffects on the spatial and temporal availability of zooplankton prey forstriped bass larvae (Martino and Houde, 2010;Wood, 2000). The stron-gest recruitments of striped bass and other anadramous fishes inChesapeake Bay occur during wet years (Martino and Houde, 2010;North and Houde, 2003; Wood, 2000). At decadal time scales,Chesapeake Bay striped bass landings and juvenile recruitments arepositively correlated with the AMO while Chesapeake Bay menhadenlandings and recruitments are negatively correlated with the AMO(Wood et al., in preparation). The AMO influences the production ofChesapeake Bay striped bass and other fishes through atmosphericforcing and weather effects on hydrographic structure, the subsequenttiming andmagnitude of zooplankton prey production, and the quanti-ty and quality of habitat (Wood et al., in preparation). Collie et al.(2008) showed that species assemblages in Narragansett Bay, RhodeIsland had shifted over the last 50 years and these shifts were stronglyassociated with the AMO. There were increases in the ratio of pelagicto demersal fish within the bay and an increase in the number ofwarm water species over time. Similarly, the species assemblage inBristol Channel, a estuarine environment in southwest England, alsochanged with SST (Genner et al., 2004).

5.3. Northwest Atlantic

In the Northwest Atlantic, the AMO has been associated with shiftsin spatial distribution of fish stocks (Nye et al., 2009), loggerhead seaturtle Caretta caretta population dynamics (Van Houtan and Halley,

2011), recruitment of marine and anadramous fishes (Gröger andFogarty, 2011) and shifts in fish assemblages (Collie et al., 2008;Lucey and Nye, 2010). Landings and population abundance are coher-ent among most stocks of Atlantic salmon and correlated with theAMO (Friedland et al., in review). Growth in northeast Atlanticsalmon stocks was highly correlated with sea surface temperaturemeasurements when post-smolts enter the marine environment(Friedland et al., 2009). Poor growth under warm conditions may bea result of temperature-induced growth reduction and/or declinesin productivity associated with the positive phase of the AMO sincethe 1990s and may be one reason that Atlantic salmon has not recov-ered despite reduced exploitation.

Much of the research relating the AMO to ecosystem change hasbeen done in the current warm phase of the AMO. However, historicalstudies indicate similar changes in species assemblages in the previ-ous warm phase of the AMO that occurred in the 1950s. Landingsof yellowtail flounder Limanda ferruginea, silver hake Merlucciusbilinearis and lobster Homarus americanus, declined in the southernparts of their ranges (south of Cape Cod), but increased in northernparts of their range (Taylor et al., 1957). Menhaden was absentfrom Massachusetts waters from 1900 to 1922, but reappeared andsupported a fishery in 1945. New reports of southern species in north-ernwaters increased aswell and these changes seemed to be associatedwithwarm temperatures (Taylor et al., 1957). After 1953, temperaturesdeclined again in this region, but southward shifts in distribution werenot documented in the same manner as the northward shifts observedduring the warming phase (Colton, 1972). A disease of the Eel grassZostera spread rapidly along the Atlantic coast of North America duringan earlier warm phase when cultural eutrophication was less of adominate driver of seagrass decline (Petersen, 1935).

Steele et al. (2007) noted that low nitrate levels associated withlower fish productivity prevailed during the 1960s followed by an in-crease in the early 1970s. These nutrient levels may have been associ-ated with climatic conditions related to NAO, particularly the effectsof Labrador Current transport. However, the AMO and its effect oncirculation patterns may have also played a role.

Several whole-ecosystem analyses have identified the AMO as animportant factor regulating ecosystem state. A multivariate analysisof the Northeast US LME found that anthropogenic factors, primarilyfishing, caused large scale changes in the ecosystem. An array of cli-mate and physical indicators were examined and revealed a transi-tion beginning in the mid-to-late 1990s, reflecting an increase in theboth the NAO and AMO along with associated temperature, riverflow, and wind stress metrics in the more northerly locations(Fig. 4). In Canadian systems similar patterns were also observed onthe Scotian Shelf (Frank et al., 2005) and a coastwide analysis oftrends in ecosystem variables revealed that fishing was the dominantpressure on the ecosystem, but that the AMO was also an importantfactor (Shackell et al., 2012). In these systems, it appears thatoverfishing triggered a cascading effect which has been amplifiedand maintained by environmental change even as fishing pressurehas been reduced on many ecosystem components. The persistentpositive AMO and NAO state in this region may explain why manyfish stocks have not recovered, while others have.

The local mechanism in the Northwest Atlantic that links AMOCand the AMO to these observed ecosystem changes is the interplaybetween warm slope water influenced by the Gulf Stream and coldLabrador slope water and the resulting spatial pattern in temperature.This process affects the entire shelf, but particularly affects nutrientand temperature regime in the Gulf of Maine. Bottom water entersthe Gulf of Maine through the narrow Northeast Channel. If thevolume of cold Labrador slope water is large relative to warm GulfStream water, bottom water temperatures in the Gulf of Maine andouter continental shelf will be relatively cold. However, when theGulf Stream is in a more northerly position, shelf waters tend to bewarmer. The exchange between these two water masses has been

110 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

shown to affect distribution of commercially important fish, abundanceof Calanus finmarchicus, and right whale calving success (Green andPershing, 2003; Nye et al., 2011).

5.4. Northeast Atlantic

Although the AMO was not formally named prior to the 1990s,historical records indicate that shifts in species distributions occurredin the 1930s in European waters during a positive phase of the AMO(Southward, 1980). Russell and Kemp (1932) reported the appear-ance of Velella, Ianthina, and Salpa fusiformis at Plymouth, UK. Whalesand dolphins were more common in Scottish water (Stephen, 1938).Strandings of whales and dolphins, uncommon in Scottish waters,became much more frequent in 1931–1933. Several Kemp's Ridley

Fig. 4. Results from PCA analysis of a Northwest Atlantic ecosystem (Northeast US LME)showing anthropogenic factors (top panel), climate stressors including the AMO (middlepanel) and ecosystem responses (bottom panel). The gray highlighted area indicates thetime period where a shift in ecosystem status occurred. Modified from Ecosystem StatusReport for the Northeast US LME (For more details see Ecosystem Assessment Program,2009).

and loggerhead turtles were caught in the English Channel perhapscarried by surface water penetrating the English Channel and/or bystrong westerly winds (Parker, 1939; Russell, 1939). During thistime, tropical tunicates and blue sharks also were more abundant innorthern waters (Farran, 1933) and a disease of the eelgrass Zosteraspread throughout Europe (Petersen, 1935). Several marine specieswere found in European waters that had never been recorded previ-ously or had been found on only one or two occasions (Stephen,1938). Dense patches of diatoms apparently affected the herringfishery by forming a barrier that herring avoided during their migration(Savage and Wimpenny, 1936). These shifts in species assemblagesthroughout the food chain were attributed to a 1.5–3 °C increase inwater temperatures in the North Sea in the early 1930s, but salinitieswere also high, indicating a change in circulation. At the time, it wassuggested that this general warming in air and water temperatureswas caused by greater northward flow of Atlantic water of southernorigin (Farran, 1933; Stephen, 1938). Anonymously high egg produc-tion of Atlantic herring and oysters were attributed to these hightemperatures. A more detailed explanation of changes in the EnglishChannel are given in Meiszkowska et al. (this edition).

Monitoring off Plymouth, England since the 1920s indicate thatduring periods of positive AMO there were high abundance of chaeto-gnaths Sagitta elegans, but low abundance of fish larvae, particularlysardine Sardina pilchardus and the opposite being true during periodsof negative AMO (Hawkins, 2003; Mieszkowska et al., this issue;Russell, 1973; Southward, 1980; Southward et al., 2005). Squid (Loligo)arrived earlier in the warmer 1950s than the cooler 1960s and 1970s(Sims et al., 2001). This phase shift was termed the “Russell Cycle”,but further monitoring found that these ecosystem phase changeswere not truly cyclical. The phase shifts associated with the “RussellCycle” appear to be driven by AMO forcing, which as described abovehas not been established as a cycle with regular oscillations either.Härkönen et al. (2006) reviewed the potential causes of two recent out-breaks of phocine distemper in the North Sea and parts of the Baltic.Whilst the specific causes are not clear, both coincide with largechanges in the cod and blue whiting stocks at a time when the AMOand NAO were in positive phases.

During the 1920s and 1930s a large scale warming event occurred inthe northern North Atlantic ocean which altered marine ecosystems ofthe Barents, Norwegian, and Icelandic Seas (Drinkwater, 2006, 2011).Warm polar water temperatures reduced the amount of sea ice andenhanced the flow of warm Atlantic water into the area. The changesin oceanographic condition are thought to have caused an increase inphytoplankton and zooplankton and actuated cascading effects onupper trophic levels of the ecosystem. The ranges of cod, haddock,and herring expanded northward while capelin and polar cod wereexcluded from the area. The phenology of migrations changed withspecies entering these waters earlier and warm-water Atlantic benthicspecies were colonized the region for some time.

In the Barents Sea, the distribution of Atlantic cod, capelin and polarcod shifted northwardwhile the distribution of haddock andNorwegianspring-spawning herring shifted eastward. The abundance of herringincreased over the prolonged warming period and their range expand-ed, supporting fisheries in both Russia and Icelandwhere these fisheriesdid not exist before (Drinkwater, 2011). During the same timeperiod, inthe Norwegian Sea, the location of the spawning sites of Atlantic codshifted to more northerly regions (Sundby and Nakken, 2008).

During positive phases of the AMO the proportion of cod spawning innorthern regions increased as did recruitmentwhile catcheswere poor infisheries that targeted cod from the southernmost spawning grounds.During negative phases of the AMO, the southernmost spawning groundshave been more important, but overall recruitment and year classstrength tended to be lower. Since 2003, spawning has beenmore activein northern Norway, in areas where it has not observed for 40 years(Sundby and Nakken, 2008). Species richness of fish in 3 regional seasin the eastern North Atlantic Ocean has changed over time (1997 to

111J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

2008) in response to higher water temperatures and the biogeographicaffinity of the species (ter Hofstede et al., 2010). In the North and CelticSeas, species richness increased due to increases in the number ofwarm-adapted Lusitanian species. In the area west of Scotland, speciesrichness decreased because the number of cold-adapted Boreal speciesdecreased. In the Irish Sea, off-shelf regions showed a progressivelylater timing in the seasonal peak of chl-a measurements moving north-wards loosely associated with the AMO (McGinty et al., 2011). Both theNAO and AMO were important to understand ecosystem changes atthese high latitudes, but the role of warm water advected from thesouth associated with the AMO into these northern areas is critical tounderstanding ecosystem changes.

The key factor to understand the physical oceanography of theNortheast Atlantic is the extent of the subpolar gyre. The subpolar gyrerotates counterclockwise and is composed of relatively cold and low sa-line water. During negatives phase of the AMO, the circulation is intenseand cold water from the gyre dominates the northern Northeast Atlantic(Benson et al., 2007; Hatun et al., 2005). Warm, saline water fromthe North Atlantic Current (NAC) originating from the Gulf Stream andEastern North AtlanticWater (ENAW) originating from the Biscay regionis prevented frommoving northward by the eastern limb of the subpolargyre. During positive phases of the AMO, the circulation of the gyreweakens and the gyre retracts, allowing the warm, saline waters of theNAC and ENAW to move northward leading to a rapid warming andsalinification of the northern Northeast Atlantic. Warm subtropicalwater is allowed to enter these areas. A recent study suggests that thesubtropical gyre contracts as a result of changing winds that can occuranywhere over the Atlantic. Changes in wind patterns or “blockingevents” vary decadaly with the AMO (Häkkinen and Rhines, 2004;Häkkinen et al., 2011). Changes in strength and extent of the subpolargyre lead to variations in the community assemblage of the NortheastAtlantic through bottom-up processes. A numerical ocean general circu-lation model links these changes in circulation with phytoplankton,zooplankton, blue whiting, and pilot whales (Hatun et al., 2009).

5.5. North Sea

Long-term shifts in species assemblages and distribution attributableto climate variability have been detected in the North Sea. An increase inthe inflow of generally warmer and more saline NAC water occurred inthe late 1980s. This caused changes at all trophic levels including in-creases in dinoflagellates and decapods (Edwards et al., 2002). As is thecase in other ecosystems, both fishing and warming have caused shiftsin abundance and distribution of fish. For example, plaice Pleuronectesplatessa showed a northward and deepening shift in distribution thatcould be attributed to climate, but not to fishing pressure, whereas soleSolea solea showed a southward shift and shallowing that was relatedto both climate change and fishing pressure (Engelhard et al., 2011).Northward latitudinal shifts in species distribution consistent with cli-mate change have been detected (Perry et al., 2005) aswell as utilizationof deeper waters by the North Sea fish assemblage driven by increasingbottomwater temperatures (Dulvy et al., 2008). The correlative relation-ships to the AMO were not developed in these studies. However, theAMO may affect the North Sea ecosystem via its influence on warmwater entering from the south through the English Channel and bywarm North Atlantic Water entering from the north. Thus, southwardshifts in distribution and invasions of new species from the north havebeen observed (Dulvy et al., 2008; Ehrich and Stransky, 2001; Perryet al., 2005).

The most recent analysis of the North Sea ecosystem covers aperiod from 1983 to 2009 and it shows a continued and intensifieddecline in the status cod and skate stocks (Kenny et al., 2009). Thesignal for these species groups were embedded in the groundfishtrend of a meta-analysis for the North Sea system (Fig. 5). However,the trends in variables showing an increase in dominance (e.g. certainpelagic fish stocks, bottom temperature and Calanus helgolandicus)

have not intensified to the same extent over the same period (Fig. 5).Indeed, the more recent data (2007–2009) indicates that the intensifi-cation of positive anomalies associated with temperature and some ofthe pelagic fish stock variables (notably herring) have actually startedto decline. It is also apparent that many variables changed state be-tween 1994 and 1996, which coincides with peaks in the positivephase of the NAO and SST in the North Sea (Fig. 5).

In an attempt to better understand the causes driving this changeduring the last 30 years, Kenny et al. (2009) used a type of ecosystemnetwork analysis (Heath, 2005; Mackinson et al., 2009). The trend inbottom-up forcing corresponds very well with the trend in the SSB ofherring for the North Sea (Fig. 6). From this we conclude that theprincipal controlling mechanism regulating the direction of the trendin the status of the North Sea herring stock is bottom-up, that is it isresponding primarily to changes in the prevailing environmental condi-tions. Clearly this trend (in terms of its rate of change) is modified byfishing pressure as is the absolute level of SSB. The relative importanceof bottom-up vs. top-down forcing is further emphasized by examininglong time-series of environment, stock and fisheries data. By comparingthe trends of North Atlantic sea surface temperature (SST-AMO),herring SSB and the herring landings there is clear correspondence be-tween the multi-decadal trend in the AMO and herring SSB, whilst thefishery appears to correspond more closely to decadal variations inherring stock yield (Fig. 6).

6. Conclusions

There is increasing evidence that the AMO is important to manytrophic levels and to overall ecosystem state, but several factors pre-vent a full understanding of how and to what extent the AMO affectsecosystems around the Atlantic basin. First, many of the observed ef-fects are based on detected changes in fish abundance surveys. Whilefish surveys provide time series longer than those typically availablefor most other taxa, they are still limited to only a few decades anddocument changes in species assemblages only for the last positivephase of the AMO. Much qualitative and some quantitative evidencesuggest that similar changes in the ecosystem occurred in warmperiods in the 1930s–50s throughout the North Atlantic. Without longtime series of multiple ecosystem metrics it is difficult to establish thelink between the AMO, changes in ecosystem state, and themore prox-imate causes of these changes. Only two studies have sufficient data toquantify population level changes occurring over more than one fullAMO cycle (Sundby and Nakken, 2008; Toresen and Østvedt, 2000).Most works only describe the recent half cycle in the AMO. Secondly,fish populations and communities have been subject to high fishingpressure in these ecosystems. Fishing pressure is a more dominantdriver in many of these ecosystems, complicating our ability to disen-tangle fishing effects from the AMO. It is likely that additional factorsincluding other climate variability patterns like the NAO interact withthe AMO making it difficult to discern the effect of the AMO itself. Assuch, a priority research area should be to coordinate research effortsto investigate the effects of other pressures (e.g. fishing) in combinationwith the AMO.

Although fish trawl surveys provide valuable time series, much ofthe published literature on the ecological effects of the AMO are focusedon autotrophs in terrestrial and marine environments. In marine sys-tems, several studies suggest that ecosystem changes related to theAMO are “bottom up” (Drinkwater, 2006, 2011). The growing body ofevidence linking large-scale climate oscillations to marine productivityemphasizes the need for greater inclusion andmore accurate represen-tation of multi-decadal climate regimes in ecosystem and global oceanmodels. The physical processes governing ocean productivity patternsoperate at multiple scales, including anthropogenic climate shifts atthe scale of centuries, decadal and multi-decadal atmospheric and oce-anic climate regimes, and inter-annual and seasonal variability. Under-standing multiple processes overlapping at different scales and how

Fig. 5. A composite and ordered plot of North Sea variable anomalies between 1983 and 2009. PC scores for each variable are given on the vertical axis and the AMO is highlighted ingray. Green indicates negative anomalies and red indicates positive anomalies. Note the intensification of negative anomalies in recent years compared to the status of the samevariables in 1983. The variables with the highest loadings and negative anomalies (top 5 variables) at the beginning of the time series were Calanus helgolandicus biomass,Horse mackerel landings, bottom silicate (μmol l−1), Atlantic herring spawning stock biomass, and average pelagic fish length. The variables with positive anomalies at the startof the time series (bottom 5 variables) were North Sea plaice CPUE, Atlantic cod landings, Whiting CPUE, North Sea plaice landings, and otter trawling effort. For more detailson methodology and variables used please see Appendix 1 and Kenny et al., 2009.

112 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

these processes influence ocean conditions and primary productivityrequires multi-decadal or centennial time series. Building, maintainingand analyzing relevant long-term time series data should be a priority(Martinez et al., 2009).

Changes in the character of several ecosystems across the Atlanticoccurred in the 1990s indicating a large-scale processwas atwork. Closerexamination of the proximate causes of these ecosystem changesemphasizes that the AMO has different modes of action throughout theAtlantic basin. For example, in the Northeast Atlantic, the primary effect

seems to stem from a change in the subpolar gyre related to more fre-quent blocking events (Häkkinen and Rhines, 2004; Häkkinen et al.,2011; Hatun et al., 2005, 2009). In the Northwest Atlantic, the primaryeffect stems from changes in the relative strength of cold Labradorslope water and warm Gulf Stream water. In terrestrial and estuarineecosystems, changes in precipitation rather than temperature associatedwith the AMO, are more important variables for ecosystem change. Al-though the AMO index is defined by SST, the AMO should not be thoughtof only as an indicator of temperature change. Thus referring to the

Fig. 6. Long-term trends in North East Atlantic spring spawning herring biomass, AMO and herring landings. The herring landings follow almost exactly the limits set by the TotalAllowable Catch (TAC).

113J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

“warm” and “cold” phase of the AMO is misleading. Rather, the AMO isan indicator of basin-wide shifts in watermasses and associated changesin atmospheric processes.

Because the AMO index represents changes in the character ofwater masses within ecosystem boundaries, we recommend thatresearchers use the AMO index as a leading indicator of large-scaleclimate variability. Once a link between ecosystem state and climatevariability is established, the more proximate mechanism/s can beidentified. When identifying proximate mechanisms, it is also impor-tant to consider the combined effect of other modes of climate vari-ability as many studies that we reviewed highlight the importanceof both the AMO and NAO (Gröger and Fogarty, 2011; Gröger et al.,2010). Furthermore, the rate of change in each ecosystem is likely afunction of whether it is a closed systems and to what degree it isclosed. Relatively closed systems vary much more rapidly than moreopen systems such as the North Sea and West of Scotland/Irelandecosystems; comparative analysis of the Baltic Sea, North Sea andIrish Seas suggest different rates of change (McGinty et al., 2011; terHofstede et al., 2010).

A key question resulting from this review of the AMO's impact onecosystem dynamics is whether the AMO can be incorporated into aresource management framework, particularly EBM. The AMO andother modes of climate variability have been successfully incorporatedinto single-species assessments. Recent modeling efforts with seaturtles (Van Houtan, 2011; Van Houtan and Halley, 2011), for example,have already developed population forecasts using oscillation indices.Furthermore, incorporating the NAO and AMO in single species assess-ment models improves the predictability of recruitment and reducesa major source of uncertainty in stock assessments (Gröger andFogarty, 2011; Gröger et al., 2010). A recent study (De Oliveira andButterworth, 2005) demonstrated that recruitment-environmentalrelationships can be useful for management even if the correlationsand associated predictability are only of moderate strength (r2>0.5).While it may not currently be feasible to incorporate environmentalindices into every single-species populationmodel, the AMO is a prom-ising metric for use in EBM.

The current body of ecological studies suggests that the AMO isassociated with multidecadal changes in ecosystem state, but canwe use the multidecadal signal of the AMO to inform EBM? If environ-mental factors can be incorporated into management a certain degreeof predictability is beneficial (Walters and Collie, 1988). Because theAMO does no fluctuate as widely as other modes of climate variabilitysuch as the NAO, the positive and negative phases of the AMO indexcan be used to predict ecosystem patterns at a coarse level. For

example, when both the AMO and NAO are in a positive phase amore rapid warming at high latitudes in the North Atlantic might beexpected and management can be adjusted depending on a species'biogeography and temperature preferences. Because the variabilityin the AMO operates on longer time scales than the NAO or annualchanges in temperature, there may be some predictability in thecycle. An analysis using global climate models indicates that AMOCand the AMO might be predictable at decadal scales (Msadek et al.,2010). Having an indicator of climate state and an understanding ofthe mechanism of its effect on different components of the ecosystemwill likely inform the management process, especially on EBM deci-sions at decadal timescales. Recent advances in simulating the dy-namics of the subpolar gyre suggests a potential for predicting thedistribution of the main faunistic zones in the northeastern Atlantica few years into the future, which might facilitate management ofthe commercially important fisheries in this region (Hatun et al.,2009).

In amanagement context, the AMOshould be viewed as an indicatorof ecosystem state. For example, in the North Sea there is a tendencytowards an ecosystem state dominated by small pelagics duringthe positive phase of the AMO,which then has a negative impact on de-mersal stocks; management could adjust fishing limits appropriately.The length of the cycle makes this hard to test statistically, but thereare mechanisms which describe how sprat and herring biomass in theNorth Sea are responding positively to increases in SST (Kenny et al.,2009). High pelagic SSB will in turn have a large negative predationpressure on the larvae of many other fish species. This multi-decadalpattern of system variability highlights some important issues relevantto fisheries and potentially ecosystemmanagement, namely that trajec-tories in ecosystem state (determined by long-term oscillations incarrying capacity) are probably the norm and they ultimately deter-mine the type and quantity of resource we can exploit. If the AMO isassociated with “bottom-up” forcing, total ecosystem production andby extension, fisheries production, may change with the phase of theAMO. It therefore follows that management approaches which assumea fixed yield or level of ecosystem service provision are inappropriate.Given that many regional management strategies now extend tocovermulti-decadal periods, knowledge of such trajectories is essential,especially as we reach the carrying capacity of many coastal systems.

Although the AMO is a mode of natural climate variability, it willinteract with and exacerbate or mitigate the effects of anthropogenicclimate change. Thus, understanding the physical aspects of thisphenomenon and its effects on ecosystems is of utmost importance.Furthermore, many of the ecosystem responses we see in the positive

114 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

phase of the AMOmay be predictors of ecosystem changes to come asglobal climate change progresses. A key component of the Earth's cli-mate system is the AMOC to which the AMO is closely associated.Global climate models predict that AMOC will weaken by 2100 byabout 25% (Meehl, 2007; Schmittner et al., 2005). For the NorthwestAtlantic, this means that the position of the north wall of the GulfStream will be in a more northerly position (Joyce and Zhang,2010), volume of Labrador slope water will lessen, and the NorthwestAtlantic ecosystems will be warmer due to the direct effects ofanthropogenic warming and to subsequent changes in circulation.With a decline in AMOC, circulation of the subpolar gyre may weakenallowing warm water from the south to influence the ecosystems ofthe Northeastern Atlantic. For this reason, continued and increasedmonitoring to improve understanding of environmental influenceson ecosystems is especially important. Such research should be trans-disciplinary, involving physical oceanographers, climate scientists, andecologists to fully understand this broad-scale phenomenon and itseffects on ecosystems. Such holistic understanding will contribute tothe effective management of natural resources through an Ecosystem-based approach.

Acknowledgements

We thank the participants of the ICES AMOworkshop (WKAMO) heldin June 2011 for thoughtful discussion that contributed to this review.Weespecially thankM. Ting, A. Gnanadesikan, R. Rykaczewski and Young-OhKwon who contributed to making an earlier version of Fig. 2. We alsothank M. Alexander, N. Shackell, and H. Walker for thoughtful reviewsof this paper. This is contribution number AED-12-034 of the AtlanticEcology Division, National Health and Environmental Effects ResearchLaboratory, Office of Research and Development, U. S. EnvironmentalProtection Agency. Although the research described in this article hasbeen funded in part by the U.S. Environmental Protection Agency, it hasnot been subjected to Agency review. Therefore, it does not necessarilyreflect the views of the Agency. This manuscript is contribution number4725 of the University of Maryland Center for Environmental Science.

Appendix 1

PC1 Variable Name PC1 Variable Name

−0.124 CalhelNS −0.024 AMO−0.124 MUR_L −0.019 Herring−0.119 Bottom SLCA (umol/l) −0.017 JOD_L−0.118 SCOPHTHALMUS RHOMBUS −0.017 Mackerel−0.110 her-SSB −0.008 MAC_L−0.105 Ave.P.Fish.L 0.007 Bottom PHOS (umol/l)−0.100 HOM_L 0.015 SOLEA SOLEA−0.099 DiaNS 0.015 her-TB−0.099 HAL_L 0.023 TRISOPTERUS ESMARKI−0.096 SCOMBER SCOMBRUS 0.026 Saithe−0.090 Sprat 0.030 euphNS−0.088 Bottom PSAL 0.031 had-TB−0.086 had-SSB 0.032 Bottom DOXY (umol/l)−0.075 Bottom NTRI (umol/l) 0.035 AMMODYTES MARINUS−0.071 meroNS 0.038 MELANOGRAMMUS AEGLEFINUS−0.071 SPRATTUS SPRATTUS 0.041 ANARHICHAS LUPUS−0.068 TRACHURUS TRACHURUS 0.042 sol-R−0.067 POLLACHIUS VIRENS 0.045 CLUPEA HARENGUS−0.066 MERLUCCIUS MERLUCCIUS 0.046 Bottom PHPH−0.062 Bottom NTRA (umol/l) 0.046 NAO−0.056 NEP_L 0.051 MERLANGIUS MERLANGUS−0.055 Bottom Temp 0.054 nop-SSB−0.045 Bottom CPHL (mg/m^3) 0.056 ple-R−0.045 LOPHIUS PISCATORIUS 0.058 Ave.D.Fish.L−0.036 PLEURONECTES PLATESSA 0.059 Bottom NTOT (umol/l)−0.029 HIPPOGLOSSUS HIPPOGLOSSUS 0.060 ple-SSB

References

Alheit, J., Pohlmann, T., Casini, M., Greve, W., Hinrichs, R., Mathis, M., O'Driscoll, K.,Vorberg, R., Wagner, C., 2012. Climate variability drives anchovies and sardinesinto the North and Baltic Seas. Prog. Oceanogr. 96, 128–139.

Andronova, N.G., Schlesinger, M.E., 2000. Causes of global temperature changes duringthe 19th and 20th centuries. Geophys. Res. Lett. 27, 2137–2140.

Beaugrand, G., Brander, K.M., Lindley, J.A., Souissi, S., Reid, P.C., 2003. Plankton effect oncod recruitment in the North Sea. Nature 426, 661–664.

Beaugrand, G., Luczak, C., Edwards, M., 2009. Rapid biogeographical plankton shifts inthe North Atlantic Ocean. Glob. Change Biol. 15, 1790–1803.

Belkin, I.M., 2009. Rapid warming of LargeMarine Ecosystems. Prog. Oceanogr. 81, 207–213.Benson, L., Petersen, K., Stein, J., 2007. Anasazi (Pre-Columbian Native-American)

migrations during the middle-12Th and late-13th centuries — were they droughtinduced? Clim. Chang. 83, 187–213.

Bostrom, B.L., Jones, T.T., Hastings, M., Jones, D.R., 2010. Behaviour and physiology: thethermal strategy of leatherback turtles. PLoS One 5, e13925.

Buchsbaum, R., Powell, J.C., 2008. Syposium review: long-term shifts in faunal assemblagesin easter North American estuaries: a review of a workshop held at the biennialmeeting of the Coastal and Estuarine Research Federation (CERF), November 2007,Providence, Rhode Island. Rev. Fish Biol. Fish. 18, 447–450.

Chavez, F.P., Ryan, J., Lluch-Cota, S.E., Ñiquen, C.M., 2003. From Anchovies to Sardinesand Back: Multidecadal Change in the Pacific Ocean. Science 299, 217–221.

Collie, J.S., Richardson, K., Steele, J.H., 2004. Regime shifts: can ecological theory illuminatethe mechanisms? Regime shifts in the ocean: Reconciling observations and theory:Progress in Oceanography, 60, pp. 281–302.

Collie, J.S., Wood, A.D., Jeffries, H.P., 2008. Long-term shifts in the species compositionof a coastal fish community. Can. J. Fish. Aquat. Sci. 65, 1352–1365.

Colton, J.B., 1972. Temperature trends and the distribution of groundfish in continentalshelf waters, Nova Scotia to Long Island. Fish. Bull. 70, 637–657.

Condron, A., DeConto, R., Bradley, R.S., Juanes, F., 2005. Multidecadal North Atlantic cli-mate variability and its effect on North American salmon abundance. Geophys. Res.Lett. 32 (23), L23703. http://dx.doi.org/10.1029/2005GL024239.

Cronin, T.M., Dwyer, G.S., Kamiya, T., Schwede, S., Willard, D.A., 2003. Medieval warmperiod, little ice age and 20th century temperature variability from ChesapeakeBay. Global and Planetary Change 36 (1), 17–29.

Cushing, D.H., 1969. The regularity of the spawning season of some fished. J. Cons. Int.Explor. Mer 33, 81–92.

De Oliveira, J.A.A., Butterworth, D.S., 2005. Limits to the use of environmental indices toreduce risk and/or increase yield in the South African anchovy fishery. Afr. J. Mar.Sci. 27, 191–203.

Delworth, T.L., Zhang, R., Mann, K.H., 2007. Decadal to centennial variability of theAtlantic from observations and models, Ocean Circulatoin: Mechanisms and Impacts.American Geophysical Union, Washington, DC.

d'Orgeville, M., Peltier, W.R., 2007. On the Pacific decadal oscillatoin and the Atlanticmultidecadal oscillation: might they be related? Geophys. Res. Lett. 34, L23705.

Drinkwater, K.F., 2006. The regime shift of the 1920s and 1930s in the North Atlantic.Prog. Oceanogr. 68, 134–151.

Drinkwater, K.F., 2011. The influence of climate variability and change on the ecosystemsof the Barents Sea and adjacent waters: review and synthesis of recent studies fromthe NESSAS Project. Prog. Oceanogr. 90, 47–61.

PC1 Variable Name PC1 Variable Name

0.060 nop-R 0.123 dinoNS0.063 calfinNS 0.126 HER_L0.065 Bottom AMON (umol/l) 0.128 cod-R0.073 ple-TB 0.129 POK_L0.073 had-R 0.131 SOL_L0.074 RAJIDAE 0.136 ANF_L0.074 her-47d3 0.136 HAD_L0.076 SPR_L 0.137 SKA_L0.078 GADUS MORHUA 0.140 PT0.079 Sole 0.142 had-340.080 SCOPHTHALMUS MAXIMUS (PSETTA MAXIMA) 0.142 NOP_L0.082 sai-3a46 0.142 nop-340.087 nop-TB 0.144 cod-SSB0.089 sol-SSB 0.145 copsNS0.090 BT 0.151 cod-TB0.091 cladNS 0.151 cod-347d0.092 Bottom TPHS (umol/l) 0.155 TUR_L0.103 WHG_L 0.155 DS0.105 sol-TB 0.155 Cod0.106 BLL_L 0.156 ple-nsea0.112 Haddock 0.157 Plaice0.116 Norway pout 0.159 CAA_L0.117 sol-nsea 0.159 COD_L0.118 her-R 0.159 Whiting0.120 MT 0.161 PLE_L0.122 HKE_L 0.161 OT

115J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

Dulvy, N.K., Rogers, S.I., Jenning, S., Stelzenmuller, V., Dye, S.R., Skjoldal, H.R., 2008.Climate change and deepening of the North Sea fish assemblage: a biotic indicatorof warming seas. J. Appl. Ecol. 45, 1029–1039.

Ecosystem Assessment Program, 2009. Ecosystem Assessment Report for the NortheastUS Continental Shelf Large Marine Ecosystem. US Department of Commerce,Northeast Fisheries Science Center Reference Document, Woods Hole, MA, p. 34.

Edwards, M., Beaugrand, G., Reid, P.C., Rowden, A.A., Jones, M.B., 2002. Ocean climateanomalies and the ecology of the North Sea. Mar. Ecol. Prog. Ser. 239, 1–10.

Ehrich, S., Stransky, C., 2001. Spatial and temporal changes in the southern species com-ponent of the North Sea bottom fish assemblages. Senckenberg. Marit. 31, 143–150.

Enfield, D.B., Mestas-Nunez, A.M., Trimble, P.J., 2001. The Atlantic multidecadal oscillationand its relation to rainfall and river flows in the continental U.S. Geophys. Res. Lett.28, 2077–2080.

Engelhard, G.H., Pinnegar, J.K., Kell, L.T., Rijnsdorp, A.D., 2011. Nine decades of NorthSea sole and plaice distribution. ICES J. Mar. Sci. 68, 1090–1104.

Farran, G.P., 1933. Nature 131.Feng, S., Hu, Q., Oglesby, R., 2011. Influence of Atlantic sea surface temperatures on

persistent drought in North America. Clim. Dyn. 37, 569–586.Frank, K.T., Petrie, B., Choi, J.S., Leggett, W.C., 2005. Trophic cascades in a formerly cod-

dominated ecosystem. Science 308, 1621–1623.Frankignoul, C., de Coetlogon, G., Joyce, T.M., Dong, S., 2001. Gulf stream variability and

ocean–atmosphere interactions. J. Phys. Oceanogr. 31, 3516–3529.Friedland, K.D., Todd, C.D., 2012. Changes in Northwest Atlantic Arctic and Subarctic

conditions and the growth response of Atlantic salmon. Polar Biol. 35, 593–609.Friedland, K.D., MacLean, J.C., Hansen, L.P., Peyronnet, A.J., Karlsson, L., Reddin, D.G.,

Maoileidigh, N.O., McCarthy, J.L., 2009. The recruitment of Atlantic salmon inEurope. ICES J. Mar. Sci. 66, 289–304.

Friedland, K.D., Shank, B.V., Todd, C.D., McGinnity, P., Nye, J., in review. Differentialresponse of continental stock complexes of Atlantic salmon (Salmo salar) to theAtlantic Multi-decadal Oscillation. J. Mar. Syst.

Fry, F.E.J., 1971. The effect of environmental factors on the physiology of fish. In: Hoar,W.S., Randall, D.J. (Eds.), Fish Physiology. Academic Press, New York, pp. 1–98.

Gaiser, E.E., Deyrup, N.D., Bachmann, R.W., Battoe, L.E., Swain, H.M., 2009. Multidecadalclimate oscillations detected in a transparency record from a subtropical Floridalake. Limnol. Oceanogr. 54, 2228–2232.

Genner, M.J., Sims, D.W., Wearmouth, V.J., Southall, E.J., Southward, A.J., Henderson,P.A., Hawkins, S.J., 2004. Regional climatic warming drives long–term communitychanges of British marine fish. Proc. R. Soc. London, Ser. B 271, 655–661.

Goldenberg, S.B., Landsea, C.W., Mestas-Nunez, A.M., Gray, W.M., 2001. The recent increasein Atlantic hurricane activity: causes and implications. Science 293, 474–479.

Gray, S.T., Graumlich, L.J., Betancourt, J.L., Pederson, G.T., 2004. A tree-ring based reconstruc-tion of the Atlantic Multidecadal Oscillation since 1567 A.D. Geophys. Res. Lett. 31.

Green, C., Pershing, A.J., 2003. The flip-side of the North Atlantic Oscillation and modalshifts in slope-water circulation patterns. Limnol. Oceanogr. 48, 319–322.

Grinnell, J., 1922. The role of the “accidental”. Auk 39, 373–380.Gröger, J.P., Fogarty, M.J., 2011. Broad-scale climate influences on cod (Gadus morhua)

recruitment on Georges Bank. ICES J. Mar. Sci. J. Cons. 68 (3), 592–602.Gröger, J.P., Kruse, G.H., Rohlf, N., 2010. Slave to the rhythm: how large-scale climate

cycles trigger herring (Clupea harengus) regeneration in the North Sea. ICES J. Mar.Sci. J. Cons. 67, 454–465.

Grossmann, I., Klotzbach, P.J., 2009. A review of North Atlantic modes of naturalvariability and their driving mechanisms. J. Geophys. Res. 114.

Häkkinen, S., Rhines, P.B., 2004. Decline of Subpolar North Atlantic Circulation Duringthe 1990s. Science 304, 555–559.

Häkkinen, S., Rhines, P.B., Worthen, D.L., 2011. Atmospheric blocking and Atlanticmultidecadal ocean variability. Science 334, 655–659.

Hare, J.A., Able, K.W., 2007. Mechanistic links between climate and fisheries along theeast coast of the United States: explaining population outbursts of Atlantic croaker(Micropogonias undulatus). Fish. Oceanogr. 16, 31–45.

Hare, J.A., Alexander, M.A., Fogarty, M.J., Williams, E.H., Scott, J.D., 2010. Forecasting thedynamics of a coastal fishery species using a coupled climate-population model.Ecol. Appl. 20, 452–464.

Härkönen, T., Dietz, R., Reijnders, P., Teilmann, J., Harding, K., Hall, A., Brasseur, S.,Siebert, U., Goodman, S.J., Jepson, P.D., Rasmussen, T.D., Thompson, P., 2006. The1988 and 2002 phocine distemper virus epidemics in European harbour seals.Dis. Aquat. Org. 68, 115–130.

Hatun, H., Sando, A.B., Drange, H., Hansen, B., Valdimarsson, H., 2005. Influence of theAtlantic subpolar gyre on the thermohaline circulation. Science 309, 1841–1844.

Hatun, H., Payne, M.R., Beaugrand, G., Reid, P.C., Sando, A.B., Drange, H., Hansen, B.,Jacobsen, J.A., Bloch, D., 2009. Large bio-geographical shifts in the north-easternAtlantic Ocean: From the subpolar gyre, via plankton, to blue whiting and pilotwhales. Prog. Oceanogr. 80, 149–162.

Hawkins, S.J., 2003. Detection of environmental change in a marine ecosystem-evidence from the western English Channel. Sci. Total. Environ. 310, 245–256.

Hawkins, S.J., Sugden, H.E., Mieszkowska, N., Moore, P.J., Poloczanska, E., Leaper, R.,Herbert, R.J.H., Genner, M.J., Moschella, P.S., Thompason, R.C., Jenkins, S.R.,Southward, A.J., Burrows, M.T., 2009. Consequences of climate-driven biodiversitychanges for ecosystem functioning of North European rocky shores. Mar. Ecol.Prog. Ser. 396, 245–259.

Heath, M.R., 2005. Changes in the structure and function of the North Sea fish foodweb,1973–2000, and the impacts of fishing and climate. ICES J. Mar. Sci. 62, 1202.

Hoar, W.S., 1953. Control and timing of fish migration. Biol. Rev. 28, 437–452.Houde, E.D., 1987. Fish early life dynamics and recruitment variability. Am. Fish. Soc.

Symp. 2, 17–29.Hu, Q., Feng, S., 2008. Variation of the North American summer monsoon regimes and

the Atlantic multidecadal oscillation. J. Clim. 21, 2371–2383.

Hubeny, J.B., King, J.W., Santos, A., 2006. Subdecadal to multidecadal cycles of LateHolocene North Atlantic climate variability preserved by estuarine fossil pigments.Geology 34, 569–772.

Hurrell, J.W., Deser, C., 2009. North Atlantic climate variability: the role of the NorthAtlantic Oscillation. J. Mar. Syst. 78, 28–41.

Hurrell, J.W., Kushnir, Y., Ottersen, G., Visbeck, M., 2003. An overview of the NorthAtlantic oscillation. The North Atlantic Oscillation: Climatic Significance andEnvironmental Impact: Geophysical Monograph, pp. 1–35.

Huss, M., Hiock, R., Bauder, A., Funk, M., 2010. 100-year mass changes in the Swiss Alpslinked to the Atlantic Multidecadal Oscillation. Geophys. Res. Lett. 37, L10501.

Joyce, T.M., Zhang, R., 2010. On the path of the Gulf Stream and the Atlantic meridionaloverturning circulation. J. Clim. 23, 3146–3154.

Joyce, T.M., Kwon, Y.-O., Yu, L., 2009. On the relationship between synoptic wintertimeatmospheric variability and path shifts in the Gulf Stream and Kuroshio Extension.J. Clim. 22, 3177–3192.

Kamenos, N.A., 2010. North Atlantic summers have warmed more than winters since1353, and the response of marine zooplankton. Proc. Natl. Acad. Sci. U. S. A. 107,22442–22447.

Kane, J., 2011. Inter-decadal variability of zooplankton abundance in the Middle AtlanticBight. J. Northwest Atl. Fish Sci. 43, 81–92.

Keller, A.A., Klein-MacPhee, G., 2000. Impact of elevated temperature on the growth,survival and trophic dynamcis of winter flounder larvae: a mesocosm study. Can.J. Fish. Aquat. Sci. 57, 2382:2392.

Kenny, A.J., Skjoldal, H.R., Engelhard, G.H., Kershaw, P.J., Reid, J.B., 2009. An integratedapproach for assessing the relative significance of human pressures and environmentalforcing on the status of Large Marine Ecosystems. Prog. Oceanogr. 81, 132–148.

Kerr, R.A., 2005. Atlantic climate pacemaker for millenia past, decades hence? Science309, 41–42.

Kilbourne, K.H., Alexander, M., Nye, J.A., in review. A Paleoclimate perspective on At-lantic Multidecadal Variability. Journal of Marine Systems.

Knight, J.R., Allan, R.J., Folland, C.K., Vellinga, M., Mann, M.E., 2005. A signature ofpersistent natural thermohaline circulation cycles in observed climate. Geophys.Res. Lett. 32, L20708.

Knudsen, M.F., Seidenkrantz, M.-S., Jacobsen, B.H., Kuijpers, A., 2011. Tracking theAtlantic Multidecadal Oscialltion through the last 8000 years. Nat. Commun. 2, 1–8.

Kwon, Y.-O., Deser, C., 2007. North Pacific decadal variability in the community climatesystem model Version 2. J. Clim. 20.

Kwon, Y.-O., Alexander, M.A., Bond, N.A., Frankignoul, C., Nakamura, H., Qiu, B.,Thompson, L., 2010. Role of the Gulf Stream and Kuroshio–Oyashio systems inlarge-scale atmospheric-ocean interaction: a review. J. Clim. 23, 3249–3281.

Li, Y.X., Yu, Z.C., Kodama, K.P., 2007. Sensitive moisture response to Holocene millenia-sclaeclimate variations in the Mid-Atlantic region, USA. Holocene 17.

Lucey, S.M., Nye, J.A., 2010. Shifting species assemblages in the northeast US continentalshelf large marine ecosystem. Mar. Ecol. Prog. Ser. 415, 23–33.

Mackinson, S., Daskalov, G., Heymans, J.J., Neira, S., Arancibia, H., Zetina-Rejón, M., Jiang,H., Cheng, H.Q., Coll, M., Arreguin-Sanchez, F., Keeble, K., Shannon, L., 2009. Whichforcing factors fit? Using ecosystem models to investigate the relative influence offishing and changes in primary productivity on the dynamics of marine ecosystems.Ecol. Model. 220, 2972–2987.

Martinez, E., Antoine, D., D'Ortenzio, F., Gentili, B., 2009. Climate-driven basin-scaledecadal oscillations of oceanic phytoplankton. Science 326, 1253–1256.

Martino, E.J., Houde, E.D., 2010. Recruitiment of striped bass in Chesapeake Bay: spatialand temporal environmental variability and availability of zooplankton prey. Mar.Ecol. Prog. Ser. 409, 213–228.

McGinty, N., Power, A.M., Johnson, M.P., 2011. Variation among northeast Atlanticregions in the responses of zooplankton to climate change: not all areas followthe same path. J. Exp. Mar. Biol. Ecol. 400, 120–131.

Meehl, G.A., 2007. Global climate projections. In: Solomon, S. (Ed.), Climate Change 2007:The Physical Science Basis. Contribution ofWorking Group I to the Fourth AssessmentReport of the Intergovernmental Panel on Climate Change (Cambridge, UK).

Mendoza, B., Velasco, V., Jáuregui, E., 2006. A study of historical droughts in SoutheasternMexico. J. Clim. 19, 2916–2934.

Mieszkowska, N., Burrows, M.T., Pannacciulli, F.G., Hawkins, S.J., (this issue). Multidecadalsignals within co-occuring intertidal barnacles Semibalanus balanoides andChthamalus spp. linked to the Atlantic Multidecadal Oscillation (this issue).

Moran, P.A.P., 1953. The statistical analysis of the Canadian lynx cycle. 2. Synchronizationand meterology. Aust. J. Zool. 1, 291–298.

Msadek, R., Dixon, K.W., Delworth, T.L., Hurlin, W., 2010. Assessing the predictability ofthe Atlantic meridional overturning circulation and associated fingerprints.Geophys. Res. Lett. 37, L19608.

Myers, R.A., 1998. When do environment-recruit correlations work? Rev. Fish Biol.Fish. 8, 285–305.

North, E., Houde, E.D., 2003. Linking ETM physics, zooplankton prey, and fish early-lifehistories to white perchMorine americana and striped bassM. saxatilis recruitmentsuccess. Mar. Ecol. Prog. Ser. 260, 219–236.

Nye, J.A., Link, J.S., Hare, J.A., Overholtz, W.J., 2009. Changing spatial distribution of fishstocks in relation to climate and population size on the Northeast United Statescontinental shelf. Mar. Ecol. Prog. Ser. 393, 111–129.

Nye, J.A., Joyce, T.M., Kwon, Y.-O., Link, J.S., 2011. Silver hake tracks changes in NorthwestAtlantic circulation. Nat. Commun. 2, 412.

O'Connor, M.P., 2010. Identifying Critical Fish Habitat and Long-term Trends in FishAbundances in Hudson River Estuary. University of Massachusett, Amherst, Amherst,MA.

Parker, H.W., 1939. Marine turtles as current indicators. Nature 143, 121.Perry, A.L., Low, P.J., Ellis, J.R., Reynolds, J.D., 2005. Climate change and distribution

shifts in marine fishes. Science 308, 1912–1915.

116 J.A. Nye et al. / Journal of Marine Systems 133 (2014) 103–116

Petersen, H.E., 1935. Preliminary Report on the disease of the eelgrass (Zostera marine L.).Rep. Dan. Biol. Stn 40, 3.

Platt, T., Fuentes-Yaco, C., Frank, K.T., 2003. Spring algal bloom and larval fish survival.Nature 423, 398–399.

Prange, H.D., 1976. Energetics of swimming of a sea turtle. J. Exp. Biol. 64, 1–12.Rodrigues, T., Grimalt, J.O., Abrantes, F.G., Flores, J.A., Lebreiro, S.M., 2009. Holocene

interdependence of changes in sea surface temperature, productivity, and fluvialinputs in the Iberian continental shelf (Tagus mud patch). Geochem. Geophys.Geosyst. 10.

Russell, F.S., 1939. Turtles in the English Channel. Nature 143, 206–207.Russell, F.S., 1973. A summary of the observations on the occurrence of planktonic

stages of fish off Plymouth, 1924–1972. J. Mar. Biol. Assoc. U. K. 53, 347–355.Russell, F.S., Kemp, S., 1932. Pelagic animals off the South-west coasts of the British

Isles. Nature 130, 664.Sanchez-Rubio, G., Perry, H.M., Biesiot, P.M., Johnson, D.R., Lipcius, R.H., 2011. Climate-

related hydrological regimes and their effects on abundance of juvenile blue crabs(Callinectes sapidus) in the northcentral Gulf of Mexico. Fish. Bull. 109, 139–146.

Savage, R.E., Wimpenny, R.S., 1936. Phytoplankton and Herring. Great Britain FisheryReports Fishery Investigations Series II: Great Britain Ministry of Agriculture andFisheries, p. 88.

Schlesinger, M.E., Ramankutty, N., 1994. An oscillation in the global climate system ofperiod 65–70 years. Nature 367, 723–726.

Schmittner, A., Latif, M., Schneider, B., 2005. Model projections of the North Atlanticthermohaline circulation for the 21st century assessed by observations. Geophys.Res. Lett. 32.

Schofield, O., Chant, R., Cahill, B., Castelao, R., Gong, D., Kahl, A., Kohut, J., Montes-Hugo,M., Ramadurai, R., Ramey, P., Yi, X., Glenn, S., 2008. The decadal view of the Mid-Atlantic Bight for the COOLroom: is our coastal system changing? Oceanography21, 108–117.

Shackell, N.L., Bundy, A., Nye, J.A., Link, J.S., 2012. Common large-scale responses toclimate and fishing across ecosystems in the northwest Atlantic. ICES J. Mar. Sci.69, 151–162.

Shen, C., Wang, W.-C., Zeng, G., 2011. Decadal variability in snow cover over the TibetanPlateau during the last two centuries. Geophys. Res. Lett. 38, L10703.

Simonti, A.L., Eastman, J.R., 2010. 2005 Caribbean mass coral bleaching event: a seasurface temperature empirical orthogonal teleconnection analysis. J. Geophys. Res.Oceans 115.

Simpson, S.D., Jennings, S., Johnson, M.P., Blanchard, J.L., Schon, P., Sims, D.W., Genner,M.J., 2011. Continental shelf-wide response of a fish assemblage to rapid warmingof the sea. Curr. Biol. 21, 1565–1570.

Sims, D.W., Genner, M.J., Southward, A.J., Hawkins, S.J., 2001. Timing of squid migrationreflects North Atlantic climate variability. Proc. R. Soc. London, Ser. B 268, 2607–2611.

Southward, A.J., 1980. The western English Channel—an inconstant ecosystem? Nature285, 361–366.

Southward, A.J., Langmead, O., Hardman-Mountford, N.J., Aiken, N.J., Boalch, G.T.,Dando, P.R., Genner, M.J., Joint, I., Kendall, M., Halliday, N.C., Harris, R.P., Leaper,R., Mieszkowska, N., Pingree, R.D., Richardson, A.J., Sims, D.W., Smith, T., Walne,A.W., Hawkins, S.J., 2005. Long-term oceanographic and ecological research inthe western English Channel. Adv. Mar. Biol. 47, 1–105.

Steele, J.H., Collie, J.S., Bisagni, J.J., Gifford, D.J., Fogarty, M.J., Link, J.S., Sullivan, B.K.,Sieracki, M.E., Beet, A.R., Mountain, D.G., Durbin, E.G., Palka, D., Stockhausen,W.T., 2007. Balancing end-to-end budgets of the Georges Bank ecosystem. Prog.Oceanogr. 74, 423–448.

Stephen, A.C., 1938. Temperature and the incidence of certain species in western Europeanwaters in 1932–1934. Anim. Ecol. 7, 125–129.

Sundby, S., Nakken, O., 2008. Spatial shifts in spawning habitats of Arcto-Norwegiancod related to multidecadal climate oscillations and climate change. ICES J. Mar.Sci. 65, 953–962.

Sutton, R.T., Hodson, D.L.R., 2005. Atlantic Ocean forcing of North American and Europeansummer climate. Science 309, 115–118.

Sutton, R.T., Hodson, D.L.R., 2007. Climate response to basin-scale warming and coolingof the North Atlantic Ocean. J. Clim. 20, 891–907.

Taylor, D.L., Collie, J.S., 2003. Effect of temperature on the functional response andforaging behavior of the sand shrimp Crangon septemspinosa preying onjuvenile winter flounder Pseudopleuronectes americanus. Mar. Ecol. Prog. Ser.263, 217–234.

Taylor, C.C., Bigelow, H.B., Graham, H.W., 1957. Climate trends and the distribution ofmarine animals in New England. U.S. Natl. Mar. Fish. Bull. 57, 293–345.

ter Hofstede, R., Hiddink, J.G., Rijnsdorp, A.D., 2010. Regional warming changes fishspecies richness in the eastern North Atlantic Ocean. Mar. Ecol. Prog. Ser. 414, 1–9.

Ting, M., Kushnir, Y., Seager, R., Li, C., 2011. Robust features of Atlantic multi-decadalvariability and its climate impacts. Geophys. Res. Lett. 38, L17705.

Tivy, A., Howell, S.E.L., Alt, B., Yackel, J.J., Carrieres, T., 2010. Origins and levels of seasonalforecast skill for sea ice in Hudson Bay using canonical correlation analysis. J. Clim. 24,1378–1395.

Toresen, R., Østvedt, O.J., 2000. Variation in abundance of Norwegian spring-spawningherring (Clupea harengus, Clupeidae) throughout the 20th century and the influenceof climatic fluctuations. Fish Fish. 1, 231–256.

Tourre, Y.M., Lluch-Cota, S.E., White, W.B., 2007. Global multi-decadal ocean climateand small-pelagic fish population. Environ. Res. Lett. 2, 034005.

Tourre, Y.M., Rousseau, D., Jarlan, L., Ladurie, E.L.L., Daux, V., 2011. Western Europeanclimate, and Pinot noir grape harvest dates in Burgundy, France, since the 17thcentury. Clim. Res. 46, 243–253.

Van Houtan, K.S., 2011. Assessing the impact of fishery actions to marine turtlepopulations in the North Pacific using classical and climate-based models. InternalReport IR-11-024. NOAA Fisheries Pacific Islands Fisheries Science Center, Honolulu,HI USA.

Van Houtan, K.S., Halley, J.M., 2011. Long-term climate forcing in loggerhead sea turtlenesting. PLoS One 6, e19043.

Vincze, M., Janosi, I.M., 2011. Is the Atlantic Multidecadal Oscillation (AMO) a statisticalphantom? Nonlinear Process. Geophys. 18, 469–475.

Walters, C.J., Collie, J.S., 1988. Is research on environmental factors useful to fisheriesmanagement? Can. J. Fish. Aquat. Sci. 45, 1848–1854.

Wanamaker, A.D., Kreutz, K.J., Schoene, B.R., Pettigrew, N., Borns, H.W., Introne, D.S.,Belknap, D., Maasch, K.A., Feindel, S., 2008. Coupled North Atlantic slope waterforcing on Gulf of Maine temperatures over the past millennium. Clim. Dyn. 31,183–194.

Wood, R.J., 2000. Synoptic Scale Climatic Forcing of Multispecies Fish RecruitmentPatterns in Chesapeake Bay, School of Marine Science. The College of Williamand Mary, Williamsburg, VA, p. 163.

Wood, R.J., Martino, E.J., Jackson, J., in preparation. Climate forcing of fisheries productionin Chesapeake Bay and the mid-Atlantic coast.

Zhang, R., Delworth, T.L., 2006. Impact of Atlantic multidecadal oscillations on India/Sahel rainfall and Atlantic hurricanes. Geophys. Res. Lett. 33, L17712.

Zhao, C., Yu, Z.C., Li, L., Bebout, G., 2010. Major shifts in multidecadal moisture variabilityin the mid-Atlantic region during the last 240 years. Geophys. Res. Lett. 37.